Friday, April 09, 2010

Angels Series Notes

Bigger Than It Seems
They are your first place Minnesota Twins.

Winning their first series of the year is exactly what you would expect from a team that is expected to challenge for the divisional crown. Because of the expectations heaped on the Twins, a triumphant first series might not seem especially important.

So let's rephrase this week's success. The Twins went into a 97-win team's stadium and took three out of four games, tallying 22 runs to the Angels' 12. And while the Angels lost John Lackey and Vladamir Guerrero from last year's AL runner-up, they also gained Joel Pineiro and Hideki Matsui to replace them, and now they have a healthy Ervin Santana. The Angels aren't just any team.

Our Twins get another very good litmus test this weekend, when they visit Chicago. We get to see if Francisco Liriano is ready for prime time. They face their likely division challengers on the road in a park where they've had little success over the last two years. And due to some messed up scheduling, the Twins should be tired as hell for today's game - usually the schedule makers would have allowed yesterday's game to be an afternoon start.

I'll be honest - I'll also be a little giddy. If they come out of this series with two wins, I'm be planning the parade before the home opener.

The Sacrifice Bunt
Lost in last night eventual 10-1 route was a somewhat controversial decision in the 7th inning which ended up being the wrong one. What we don't know is if there was a right one. With runners on 1st and 2nd base and no outs, and with the score only 3-1, Denard Span bunted to move the runners and gave himself up at first.

There are generally some pretty strong feelings on both sides of this decision. The first is semi-old-school: this is a great play, in part because it is automatically smart and sacrifices an at-bat for the good of the team. The second is more sabremetric, which teaches that most of the time bunting is stupid and costs a team runs.

In this case, neither is right. The baseline for this decision can also be evaluated two ways: which strategy is likely to score the most run vs. which strategy is most likely to result in a win. And in both cases, the answer is essentially a push. There is almost no difference is how many runs an average team playing against an average team will score if they bunt those runners over as opposed to swing away. And there is almost no change in the likelihood of winning the game.

But the key to that last paragraph was the italicized word "average". That is nothing more than a baseline evaluation. What really needs to be evaluated is how these teams differ from an average team.

The batter who was up at the plate, is decidedly above average, especially when it comes to getting on base. Above all, Span wants to avoid the double play, and with his speed, he's in a pretty good position to do so. On the other hand, he was 0-3 with three groundouts to the pitcher and second baseman last night. He was also involved in an injury break the previous night that required icing. He could very well be suffering a short-term decline in performance from it, and his manager (or Span) might have noticed that.

The batter following Span is Orlando Hudson, the team's #2 hitter, who is also an above average hitter. If Span bunts, it's his job to get a run home via a fly ball or smoked ground ball. In this case, he grounded out to a drawn in infield, and the runner from third could not score.

Finally, the sacrifice virtually ensured that Joe Mauer would come to bat. Mauer is obviously WAY above the average player, exactly who any manager would want to have up with two outs and a runner in scoring position. Again, it didn't work out, as he grounded out to third base and the Twins ended up with zero runs.

Obviously, that's just half the equation. We could also talk about the Angels pitcher, and to what extent he induced groundballs or strikeouts. Or evaluate their infield, which looked pretty strong in three of the four spots yesterday.

If the Twins have a weak hitter up in that position, followed by two strong hitters following, bunting is likely a winner, no question. If they have good hitter, followed by two weaker hitters, it's likely a mistake. In this case, they had three good hitters, which makes it less clear. I'm swayed a bit by Span's earlier at-bats that maybe he wasn't totally healthy last night. On the other hand, Hudson didn't impress offensively or defensively this series, though I expect him to be better over the season.

Whatever the smart decision was, this one didn't score any runs, so I think we can safely say it was the wrong one. What we don't know is if Span swinging away would have turned out any better. There may not have been a right decision here.


I don't know why we feel like we need to wait for a pennant race to come together as a community and get excited about this team. For all we know, this weekend might BE the pennant race. If the Twins sweep the Sox and take a 4.5 game lead over them in the first week....

OK, I'm being giddy again. But you can experience that giddiness and that sense of community with me tomorrow at Major's in Blaine for our Twins Viewing Party. It's a day game, so it starts at 12:05, though I would suggest getting there a little earlier for seats. $2 pints, two-for-one appetizer, a raffle for Twins tix, and lots and lots of Twins talk. I hope you can make it.


Joe said...

Love the analysis, John. Keep up the good work. I'm giddy too, for a significant portion of the year last year, before the Yanks started coming together, the Angels were considered the favorites to win the AL, and we just smoked them. We may not be hitting for average this year, but definitely for power, so you might want to say, "this isn't your Daddy's Minnesota Twins."

Anonymous said...

>>Whatever the smart decision was, this one didn't score any runs, so I think we can safely say it was the wrong one... There may not have been a right decision here.

That makes no sense, and the reason is that you're confusing right and wrong (or good and bad) decisions with good and bad outcomes. We don't have all the information to judge whether the decision was good or bad, for the reasons you mentioned. But if we sat down with everyone involved in the decision and got all the information they had access to at the time, then in principle we could determine whether the decision was the right / smart one, independent of the particular result that followed.

David said...

Funny, I was just reading about sacrifices in The Book this morning. Using The Book as a guide, the bunt seemed like a pretty bad strategy. First, you've got a really high scoring environment with the Twins and Angels, so single runs are not at such a premium. However, one single run at that point would have made a large difference in win expectancy, and since it was the seventh inning, maybe it wasn't quite so bad. However, runners on 2 and 3 with one out has a lower run expectancy than does runners on 1 and 2 with no outs.

The other problem is that the sacrifice didn't leverage the ability of the following batters to walk. If Hudson or Mauer walked, it would have made little difference because first base was open anyway. I say let Span swing away.

John said...

David, the run expectancy does "drop", but so little as to make it essentially meaningless. And win expectancy actually increases slightly as you can see on the review of the game. Span is actually awarded WPA for that bunt, although it's just .0001.

And I would also argue your contention that this is a high run-scoring environment. Both the Twins and Angels bullpens have been very good - up to that point the Twins hadn't scored a single run off the Angels bullpen this series.

Anonymous, I generally evaluate my decisions based on their results. That might not be entirely fair, so I cut myself some slack, but I find that's the best way to evaluate decisions.

Jack Ungerleider said...

I thought I was experiencing a little reverse deja vu this week. So I looked it up on Baseball Reference. Two years ago the Twins opened at home against the Los Angles Angels of Anaheim. The one the first game behind their "ace" Livan Hernandez. Then went on to loose the next three games. (For the record LA won 100 games that year. I'm not suggesting...)

David said...

Fair point, John. My point about high run scoring environment, though, is more general (which maybe is a flaw). But, just generally, the Twins and Angels have offenses capable of scoring a good number of runs at once, but you're right about the pitching and bullpen mitigating that. I'm not sure that the pitching is so effective as to make it a low-scoring environment, but you might be right.

Regarding the win versus run expectancy, if the run expectancy is essentially meaningless, is it that much less than the .0001 WPA?

I realize that using WE/RE isn't perfect, given that it's broadly based on averages across the league, and we're talking about a specific instance. I'm just not sure that given the particulars (mostly the quality of the hitters coming up) made that sacrifice a good move.

John said...

Winning 3 of 4 on the road is a great way to start, no doubt about that. I must say though that the Angels do not look like a very scary team. They have pretty good pitching, sort of like the Twins- deep but without an ace at this point. But I don't think they are going to score too many runs. Wood looked awful and a lot of their hitters have on-base issues.

Anonymous said...

It's me again, John. I'm shocked that you're missing something that seems like it would be obvious to any analytical baseball fan, so let me try my point one more time.

When there are critical uncontrollable and unpredictable factors involved in a particular outcome--like, oh I dunno, maybe what type of pitch is thrown and how well (!)--then judging a decision on the basis of a single result is silly. You might as well predict the Twins' final record for the year on the basis of whether they win or lose the first game. In either case, you just don't have enough data.

If we could do a controlled experiment in which Span tries that same bunt a few dozen or hundred times, letting all the random factors average out, then sure, we could evaluate the decision on the basis of the average result. Failing that, we have to look backward for relevant data--again, either in the stat books, or in the intuition of old baseball hands who can implicitly take account of specific factors that the stat books cannot.

I'll give you this much: if we didn't have much prior data on Span's ability as a bunter (I dunno if we do or not--this is just a hypothetical), then this particular outcome could be a significant addition to the knowledge base on which we could evaluate future decisions. But it still wouldn't be useful in judging the quality of the decision made by Span or whomever this time, with the information that was available then.

Hope that makes more sense now. In any case, I'm looking forward to many great columns from you in what I hope will be a great Twins season!


Burroughs PTA said...

Oh, I'm not saying that WE trumps RE. I'm saying that since they are both negligible. Much bigger factors, such as the players involved, should be used over the charts. Unfortunately, that doesn't clarify things either, since that wasn't clear cut in my mind. The point of the story was that you can debate this one to death - there is no obvious answer. For the record, if I had to choose, I probably would've chosen Span bunting.

And Perry, I appreciate your take, and my reply before was probably too curt. I see that we would like to view them as reasonably independent. After all, it's not reasonable to suggest that Span or Gardy could have known that Hudson's hard grounder would go right at Kendrick. Like I said, I would've probably done the same thing.

But ultimately, if I would have made that decision, I would've judged it as "wrong", because it didn't work. I'm not sure the other way would've been "right", because I'm not sure it would've worked. But I know the path I took didn't take me to my goal.

John said...

Oops, it looks like that last comment was under TVOR's login. For the record, it was the Twins Geek. :-)

TT said...

"I realize that using WE/RE isn't perfect, given that it's broadly based on averages across the league, and we're talking about a specific instance"

Isn't perfect is probably underestimating how unimportant it is.

Just imagine taking the average number of home runs that are hit by major league players and then applying it to a specific plate appearance, whether it is Justin Morneau or Nick Punto hitting. They aren't really even a starting point for the discussion.

They often get used in places where the differences aren't so stark and clearcut. But that doesn't mean the range of outcomes is any narrower. The use of the sacrifice bunt, as you can see here, is one of those.

It also demonstrates a larger problem. In order to determine probabilities you not only need enough data, you need random data. There is very little random data available in baseball.

walter hanson said...

Okay does the fact that the Twins in back to back games after they loaded the bases with no outs and failing to score to make this decision mute.

Look when you have the lead you play to add a run. Furthermore the goal was to allow at least one of the two big M and M guys a chance to swing the bat with two guys in scoring position. It was a good call to do the bunt. Especially since Spann who is a good bunter has a chance to lay it down for a hit.

As for the road trip taking three out of four from the Angels I'll take. We lost the season series last year despite the fact that we swept them in early April when their pitching was screwed up. This year we got three wins when you can argue they had their team healthy and ready.

Walter Hanson
Minneapolis, MN